Requirements: Linux or MacOS (use Docker or WSL to run in Windows), Julia version 1.6 or greater with CPLEX properly installed (see instructions below), and a VRPSolver-library binary file that can be downloaded from here.
CPLEX installation on Julia: The environment variable CPLEX_STUDIO_BINARIES should be set with the path to the CPLEX binaries folder before running this demo for the first time. To test this installation, please enter the Julia REPL (type julia in a terminal window) and type ] (without Enter), and then add CPLEX, and build. If no error occurs, you are ready.
Running the demo: The environment variable BAPCOD_RCSP_LIB should be set with the path to the VRPSolver-library binary file (including its name) whenever you run this demo. Then, inside the directory where you cloned this repository, just type julia src/run.jl data/A/A-n37-k6.vrp -m 6 -M 6 -u 950 to solve a small classical CVRP instance providing an initial upper bound.
Reporting issues: This is an ongoing work. Even the original VRPSolver interface is not complete. I am waiting for nice examples from you to implement and test the features that are missing and fix bugs. So, when reporting bugs, please choose examples as small as possible, provide the necessary data to reproduce the issue, and describe the expected behavior (the optimal solution and its cost, for example).
Acknowledgements: This demo package uses the Coluna package which implements a full Branch-and-Price algorithm in Julia, and is provided by AtOptima. The current version uses only the Julia interface provided by Coluna as a wrapper to VRPSolver but the full integration with Coluna is a possible future work.